Model characterization in measurements of environmental pollutants via data correlation of sensor outputs

The aim of this paper is to develop a simple air pollutant model for the analysis and the characterization of environmental data, acquired by means of multisensor monitoring system and elaborated by suitable software agents. Modeling is basically important in order to validate measured data and to forecast the time-varying behavior of contaminating substances. The proposed model is initially based on the individuation of possible correlations existing between some pollutants. Afterward, to increase the accuracy of estimated values, the influence of meteorological quantities is taken into account to improve significantly the so-obtained model. Finally, some information about reliability degree of estimate is provided.